Assessing Data Visualization Literacy: Design Implementation and Analysis of a Comprehensive Test

Data Visualization Literacy involves recognizing a given chart, reading it correctly, and extracting information from it.Familiarity with a specific type of visualization does not imply that the person can read or interpret it correctly.Data visualization researchers have attempted to explore and promote solutions to support data visualization comprehension activities.Studies address developing and applying tests to assess literacy, understanding how analysts interpret visualizations and how unfamiliar visualizations are taught, and even identifying the cognitive activities involved in creating visualizations.However, we have not found comprehensive studies relating to these different aspects, which we believe are essential for teaching and learning about data visualizations and data analysis tasks.This paper presents our procedure for designing a new visualization literacy assessment.We devised a test that covered 15 different visualizations and applied it with 68 participants.We analyzed each item, assessing their difficulty level and discrimination.After removing eight items that did not discriminate well, our final test ended with 37 items, all with fair or good discrimination.It means that the test should genuinely represent the test takers' learning ability, i.e., the items can discriminate well between the high and low-performing groups.